[PATCH synaptics] Add synaptics orientation support

Peter Hutterer peter.hutterer at who-t.net
Tue Mar 22 22:17:10 PDT 2011

On Mon, Mar 21, 2011 at 01:51:35PM +0200, Aapo Rantalainen wrote:
> > On Fri, Mar 18, 2011 at 04:26:46PM +1000, Peter Hutterer wrote:
> > Hi Aapo,
> >
> > I noticed this patch in the synaptics repo today. Unfortunately, it needs
> > a bit more work, so I've reverted it for now. Please find my comments
> > inline.
> > fwiw, patches to synatics should go to the xorg-devel list first for public
> > review.
> Hi, this patch is extended (=longer) version of my patch, so I hope
> Christoph would comment this too.
> I made this patch (three years ago) because I thought there was need
> for this missing feature.
> User scenario:
> User has laptop with touchpad. User rotates laptop (physical device)
> to portrait. Display is rotated, automatically or manually.
> Then user move finger on touchpad from top to bottom (real life coordinate)
> ( http://forum.eeeuser.com/viewtopic.php?pid=140403 )
> Expected:
> Mouse cursor is moving from top to bottom on the screen.
> http://www.youtube.com/watch?v=VOQds5DfC60
> Actual happens:
> Mouse cursor is moving horizontally (left or right).
> *
> Today I know that wacom has same kind of feature, but I think only
> drawing board is then rotated, not the display. (So it is not good
> idea that display always know it's orientation and forces axes to all
> interface device).
> What is then common for wacom/synaptics/others? Is there a point to
> tune this synaptics-patch or is this doable on server level? (Or some
> other better way?)

(see my other email, just one point about wacom devices)
wacom devices only rotate the input device. if they are mounted onto actual
screens like TabletPC's ISDV4 devices, then the screen rotation and the
input device rotation must be triggered independently since the wacom driver
doesn't know about output rotation.

> > why not use degrees here?
> > "left" doesn't have a clear meaning. -> "90 degrees clockwise"
> > I'm more of a fan of the  "rotation",
> > just because xrandr uses 0-3 doesn't make it a good idea ;)
> If this patch is only for touchpad, is there really use for arbitrary degrees?

mostly to unify the various rotation options, so we don't have different
values for each driver. it's easy enough to restrict the valid values to
0, 90, 180, 270.


More information about the xorg-devel mailing list